INRIA a CCSD electronic archive server
Not a member yet
122212 research outputs found
Sort by
Imputing missing minimum inhibitory concentration (MIC) values for Pseudomonas aeruginosa strains with a Denoising AutoEncoder
No full textAbstract Pseudomonas aeruginosa is a problematic pathogen with complex antibiotic resistance patterns. In clinical practice, minimum inhibitory concentration (MIC) tests typically focus on a limited subset of antibiotics, hindering a comprehensive assessment of a strain’s resistance profile. Here, we introduce MICFiller, a Denoising AutoEncoder (DAE) model designed to impute missing MIC values for 14 antibiotics in Pseudomonas aeruginosa within a specific dilution range by leveraging known MIC measurements for other antibiotics in the same strain. We evaluated the performance of DAE against two other commonly used methods: Multiple Imputation by Chained Equations (MICE) and simple median imputation. The DAE achieved the highest balanced 1-tier accuracy for most antibiotics, with performance closely matching that of MICE. MICFiller is freely accessible through a user-friendly web interface at http://iorgalab.org:4567/micfiller , offering clinicians a more complete view of a strain’s antibiotic resistance profile.Pseudomonas aeruginosa is a problematic pathogen with complex antibiotic resistance patterns. In clinical practice, minimum inhibitory concentration (MIC) tests typically focus on a limited subset of antibiotics, hindering a comprehensive assessment of a strain’s resistance profile. Here, we introduce MICFiller, a Denoising AutoEncoder (DAE) model designed to impute missing MIC values for 14 antibiotics in Pseudomonas aeruginosa within a specific dilution range by leveraging known MIC measurements for other antibiotics in the same strain. We evaluated the performance of DAE against two other commonly used methods: Multiple Imputation by Chained Equations (MICE) and simple median imputation. The DAE achieved the highest balanced 1-tier accuracy for most antibiotics, with performance closely matching that of MICE. MICFiller is freely accessible through a user-friendly web interface at http://iorgalab.org:4567/micfiller , offering clinicians a more complete view of a strain’s antibiotic resistance profile
Data efficient muscle parameter estimation: Application to the Human Upper Limb
No full textInternational audienceNumerical simulations of musculoskeletalmodels are widely used to study and predicthuman movement. To be viable, these modelsneed to be personalized to each subject. A keychallenge lies in understanding how specificmovements influence the observability ofmuscle parameters to optimize modelcalibration experimental protocols by providingtrajectories which excite these parameters,reducing experiment duration and complexity.In this article we use an open source motiondataset of the upper limb to explore data-efficient muscle parameter estimation of theupper limb
Online Stochastic Matching: A Polytope Perspective
No full textStochastic dynamic matching problems have recently gained attention in the stochastic-modeling community due to their diverse applications, such as supply-chain management and kidney exchange programs. In this paper, we study a matching problem where items of different classes arrive according to independent Poisson processes. Unmatched items are stored in a queue, and compatibility between items is represented by a simple graph, where items can be matched if their classes are connected.We analyze matching policies in terms of stability, delay, and long-term matching rate optimization. Our approach relies on the conservation equation, which ensures a balance between arrivals and departures in any stable system. Our main contributions are as follows.We establish a link between the existence of stable policies, the dimensionality of the solution set of the conservation equation, and the compatibility graph's structure.We describe the convex polytope formed by non-negative solutions to the conservation equation, and we design policies that can achieve or closely approximate the vertices of this polytope.Lastly, we discuss potential extensions of our results beyond the main assumptions of this paper
Aevol-9: A simulation platform to decipher the evolution of genome architecture
No full textAevol is a forward-in-time simulator of genome architecture. It simulates a population of individuals, each with an explicit genome whose sequence and architecture can be modified by various mutational operators, including substitutions, indels, and large-scale chromosomal rearrangements. This enables performing in silico experiments to decipher the effects of evolutionary conditions ( e . g . population size, selection strength, mutation rates, and biases) on genome organization
Outdoor Hybrid Solar Road Demonstrator Monitoring Using Infrared Thermography with Embedded Local Probes for Energy Harvesting Performance Evaluation
No full textInternational audienceThis study investigates, in a natural environment, the thermal behavior of an innovative pavement system with thermal and solar energy collection functionalities. The whole structure is continuously monitored using temperature and heat flux sensor probes integrated inside the structure. Local weather conditions are also monitored. Infrared Thermography is used as a complementary non-invasive technique to monitor temperature surface distributions with time and assess the efficiency of heat transfer within the pavement structure. All sensors are connected to a newly developed platform that centralized data access, visualization, and storage, enabling seamless management and user interactions. The obtained results are presented and discussed
Advancing Environmental Hazard Resilience through Transdisciplinary Education
No full textInternational audienc
UNIF 2025
No full textInternational audienceThis collection contains the contributions presented at the 39th International Workshop on Unification (UNIF 2025). UNIF 2025 was held in affiliation with the 10th International Conference on Formal Structures for Computation and Deduction (FSCD 2025). The workshop took place in Birmingham, United Kingdom, on July 14, 2025.Unification addresses the problem of making two given terms equal — either syntactically or modulo an equational theory. It is a fundamental process in various areas of computer science, including automated reasoning, term rewriting, logic programming, natural language processing, program analysis, knowledge representation, type theory, and more.The International Workshop on Unification (UNIF) serves as a forum for researchers in unification theory and related fields to present recent (even unfinished) work and to discuss new ideas and emerging trends. It also provides an excellent opportunity for students, young researchers, and scientists working in related areas to gain an overview of the current state of the art in unification theory. For more information about the workshop's history, please visit the UNIF homepage.The 39th edition consists of seven papers that were selected and accepted by the UNIF 2025 Program Committee. Each submission was independently evaluated and reviewed by three reviewers. UNIF 2025 also had two invited speakers: David M. Cerna (Dynatrace Research, Czech Academy of Sciences) and Oliver Fernández Gil (TU Dresden, Germany)
Wasserstein Convergence of Critically Damped Langevin Diffusions
No full textInternational audienceScore-based Generative Models (SGMs) have achieved impressive performance in data generation across a wide range of applications and benefit from strong theoretical guarantees. Recently, methods inspired by statistical mechanics, in particular, Hamiltonian dynamics, have introduced Critically-damped Langevin Diffusions (CLDs), which define diffusion processes on extended spaces by coupling the data with auxiliary variables. These approaches, along with their associated score-matching and sampling procedures, have been shown to outperform standard diffusion-based samplers numerically. In this paper, we analyze a generalized dynamic that extends classical CLDs by introducing an additional hyperparameter controlling the noise applied to the data coordinate, thereby better exploiting the extended space. We further derive a novel upper bound on the sampling error of CLD-based generative models in the Wasserstein metric. This additional hyperparameter influences the smoothness of sample paths, and our discretization error analysis provides practical guidance for its tuning, leading to improved sampling performance
Guess Future Anomalies from Normalcy: Forecasting Abnormal Behavior in Real-World Videos
No full textInternational audienceForecasting Abnormal Human Behavior (AHB) aims to predict unusual behavior in advance by analyzing early patterns of normal human interactions. Unlike typical action prediction methods, this task focuses on observing only normal interactions to predict both, short and long term future abnormal behavior. Despite its affirmative impact on society, AHB prediction remains under-explored in current research. This is primarily due to the challenges involved in anticipating complex human behaviors and interactions with surrounding agents in real-world situations. Further, there exists an underlying uncertainty between the early normal patterns and the future abnormal behaviour, thereby making the prediction harder. To address these challenges, we introduce a novel transformer model that improves early interaction modeling by accounting for uncertainties in both, observations and future outcomes. To the best of our knowledge, we are the first to explore the task. Therefore, we present a new comprehensive dataset referred to as "AHB-F" † , which features real-world scenarios with complex human interactions. The AHB-F has a deterministic evaluation protocol that ensures only normal frames to be observed for long-and short-term future prediction. We extensively evaluate and compare competitive action anticipation methods on our benchmark. Our results show that our method consistently outperforms existing action anticipation approaches, both in quantitative and qualitative evaluations
Expected Length of the Euclidean Minimum Spanning Tree and 1-norms of Chromatic Persistence Diagrams in the Plane
No full textxLet be the constant such that the expected length of the Euclidean minimum spanning tree of random points in the unit square is in the limit, when goes to infinity. We improve the prior best lower bound of by Avram and Bertsimas to . The proof is a by-product of studying the persistent homology of randomly -colored point sets. Specifically, we consider the filtration induced by the inclusions of the two mono-chromatic sublevel sets of the Euclidean distance function into the bi-chromatic sublevel set of that function. Assigning colors randomly, and with equal probability, we show that the expected -norm of each chromatic persistence diagram is a constant times in the limit, and we determine the constant in terms of and another constant, , which arises for a novel type of Euclidean minimum spanning tree of -colored point sets